Journal of Geotechnical and Geoenvironmental Engineering

The roles of particle breakage and drainage conditions on the quasi-static compression response of sand were evaluated by comparing the results from drained and undrained isotropic compression tests on dry and saturated specimens up to a mean total stress of 160 MPa. For dry sand specimens, the compression curves from drained and undrained tests were similar because of the high compressibility of air. The isotropic compression curves of the dry sand specimens at mean stresses greater than 30 MPa reflect a transition toward void closure, reaching a minimum void ratio of 0.04 at 160 MPa. Dry sand specimens with different initial relative densities showed similar behavior during isotropic compression in drained conditions for mean stresses greater than approximately 30 MPa. As expected, saturated sand specimens tested under undrained conditions showed a much stiffer response than in drained conditions, with a bulk modulus greater than that of water. Increasing trends in particle breakage quantified using breakage factors from the literature with increasing mean stress were observed for the dry sand specimens, but negligible particle breakage was observed for the saturated sand specimens tested in undrained conditions. To highlight the linkage between particle breakage and transition to void closure at high mean effective stress, an empirical relationship was developed using a breakage factor from the literature to match the experimentally-derived compression curves of dry sands under drained conditions.

Knowledge of erodibility is essential to the analysis of problems related to soil erosion including bridge scour, embankment and floodwall overtopping, dam spillway erosion, stream stability, and meander migration. Many investigators have attempted to classify erodibility according to basic engineering soil properties such as Atterberg limits, unit weight, water content, and shear strength. In general, the more successful of these attempts were those that studied the erodibility of coarse-grained materials due to the fact that the erosion resistance of such soils is primarily influenced by gravity and friction, and not the more complex electrostatic and electromagnetic forces that influence fine-grained soils behavior. This paper describes a way of assessing erodibility by presenting erodibility classification charts and erosion threshold charts for different soil types. These charts are developed from soil erosion test data as well as related experience of the authors and are meant to be used by practicing engineers as a first-order estimate in their erosion analyses. The paper also presents nonlinear models correlating soil erodibility parameters to the soil geotechnical properties. A practical example of the use of the charts is given to illustrate their application in practice.

The liquefaction potential index (LPI) has increasingly been used for assessing liquefaction hazards induced by earthquake shaking. The LPI allows an integral number of soil horizons to be evaluated for liquefaction potential. Considerable uncertainty exists in cases where the LPI thresholds for classifying liquefaction hazards statistically rely on the data sets and/or the framework. This study explored what the appropriate LPI classification thresholds should be and developed a liquefaction reduction number ($R_L$) to better account for the impact of nonliquefiable layers. The LPI and $R_L$ values were then analyzed at the standard penetration test (SPT) data sites reported in case studies. The results suggested that most SPT-based LPI thresholds should be modified upward in comparison with Iwasaki’s original scale. These corrections are of particular importance when clay minerals are part of the soil mixture, rather than simply using the minus 200-sieve fraction as the liquefaction susceptibility criterion. Nonliquefied sites appear to exhibit higher $R_L$ values, suggesting that the $R_L$ could supplement the LPI as a valuable proxy for liquefaction hazard assessments. Probabilistic liquefaction hazard assessments could be estimated using a logistic regression of the LPI and/or $R_L$ values.

Highly sensitive, low-saline clays, termed quick clays, represent a risk for large landslides. An in situ experiment was conducted in which salt wells filled with potassium chloride were installed to study how salt wells can be used to reduce landslide risk. Salt-plume migration and clay volume surrounding the salt wells were investigated by resistivity cone penetration testing (RCPTu), piston samples, conductivity divers, and groundwater samples. Correlating geotechnical properties with pore-water compositions, the remolded shear strength ($c_{ur}$) was improved beyond 1 kPa when the ratio of the sum of potassium, magnesium, and calcium to the major cations exceeded 20%. The value of $c_{ur}$ increased to approximately 4–8 kPa at a distance of 0.5 m from the wells within 3 years, and the liquidity index (LI) decreased from more than 3.4 to less than 1.2. A minimum diameter of 1.5 m around the wells was stabilized within three years. It is recommended that salt wells be installed at a center-to-center distance of 1.5–2.0 m. The improved postfailure properties ($c_{ur}$ and Atterberg limits) are considered permanent in an engineering timescale because of a lasting pore-water composition that inhibits development of high sensitivity. Salt wells can be installed without substantial disturbance of the soil and can be used as landslide mitigation in quick-clay areas.

While significant emphasis has been placed on the quantification of soil–geosynthetic properties under failure conditions, studies of properties that are suitable for characterizing this interaction under serviceability conditions have been limited. Also, most geosynthetic properties are currently defined in isolation rather than under the confinement of soil. The purpose of this study is to develop a soil–geosynthetic interaction framework that, with a single and repeatable parameter, can capture the stiffness of a soil–geosynthetic composite under small displacements. The soil–geosynthetic interaction model developed in this study involves well-established force equilibrium differential equations. However, the constitutive relationships and boundary conditions were specifically selected so that the model results in a closed-form analytical solution. Since the analytical solution involves a single parameter, its use may be particularly suitable for specifications and the design of structures such as stabilized roadways. This parameter, referred to as the stiffness of the soil–geosynthetic composite, or $K_{\text{SGC}}$, captures both the tensile characteristics of the geosynthetic and the shear behavior of the soil–geosynthetic interface. Experimental procedures to quantify $K_{\text{SGC}}$ were developed as part of this study. The results of a pilot experimental program, conducted using tailor-made soil–geosynthetic interaction equipment, are presented in the paper. These results confirm the suitability of the assumptions and outcomes of the model. A companion paper provides the results of a comprehensive experimental program with particular emphasis on the evaluation of the repeatability of the results and on the sensitivity of the assumptions and outcomes of the model to variables that impact $K_{\text{SGC}}$.

While most soil–geosynthetic interaction models have focused on the characterization of failure conditions, little emphasis has been placed on models and parameters suitable for characterizing the stiffness of soil–geosynthetic systems. In the companion paper, a soil–geosynthetic interaction parameter ($K_{\text{SGC}}$) was developed that captures the stiffness of a soil–geosynthetic composite under small displacements. This included validation of the suitability of the assumptions and outcomes of the model for a specific set of materials and testing conditions. This paper presents the results of a comprehensive experimental program that allows the suitability of the model to be generalized for a wider range of materials and testing conditions. An initial test series was conducted using large-scale soil–geosynthetic interaction test equipment to evaluate the repeatability of the experimental results. A comparison of the test results from this series, as well as an assessment of an extensive database on the expected variability of soil and geosynthetic properties, revealed that the coefficient of variation of the model parameters was acceptable and well within the typical range of similar geotechnical and geosynthetic properties. Results from additional test series confirmed the linearity and uniqueness of the relationship between the geosynthetic unit tension squared and corresponding displacements, which are the key features of the proposed model. These tests were conducted under various conditions using different geosynthetic and backfill materials. Results also showed that the constitutive relationships adopted in the model were adequate for the extended range of confining pressures, geosynthetic lengths, geosynthetic types, and backfill soil types adopted in the study. The consistency of the results obtained in the experimental testing program underscores the suitability of the proposed $K_{\text{SGC}}$ parameter as a basis for the evaluation of soil–geosynthetic interactions under small displacements.

This is the written version of the H. Bolton Seed Medal Award Lecture presented at the ASCE Geotechnical and Structural Engineering Congress in Phoenix, Arizona, February 14–17, 2016. It discusses five recent findings on liquefaction triggering of clean and silty sands during earthquakes. Tools ranging from case history analysis to centrifuge tests were used in the corresponding studies. The findings are (1) pore pressure ratio during earthquakes is more uniquely correlated to cyclic shear strain, ${\gamma }_c$, than to cyclic stress ratio (CSR); (2) current penetration and shear wave velocity ($V_s$) charts are associated with small to moderate cyclic strains triggering liquefaction, ${\gamma }_{cl}$, that range from ${\gamma }_{cl}\approx 0.03\%$ to ${\gamma }_{cl}\approx 0.3\%$ depending on soil characteristics and earthquake magnitude; (3) for recent uncompacted clean and silty sand fills that have not been significantly preshaken, such as those in the San Francisco Bay Area of California, and a Richter scale magnitude $M_w=7.5$, triggering occurs at ${\gamma }_{cl}\approx 0.03\%$; (4) for the heavily preshaken, geologically recent natural silty sands in the Imperial Valley of California, ${\gamma }_{cl}\approx 0.1–0.2\%$ with a liquefaction resistance that is twice as big despite the fact that some of these sands were deposited as recently as the uncompacted fills in San Francisco; and (5) the static cone penetration resistance (CPT) tip penetration resistance of a clean sand is more sensitive to preshaking than $V_s$, with the CPT capturing better the increased liquefaction resistance due to preshaking.

Tire-derived aggregate (TDA) has been used widely in civil engineering applications such as highway embankments, light rail foundations, landslide repairs, and retaining walls as both a recycled material and a lightweight fill. Although the shearing properties of certain types of TDA have been studied, there is still a need for representative and reliable properties of TDA with large particles, such as Type B TDA with particle sizes ranging from 150 to 300 mm. Direct shear tests were performed on Type B TDA using a new large-scale shearing device to measure properties governing internal shear strength as well as interface shear strength against concrete. The internal failure envelope is nonlinear, with a secant friction angle decreasing from 39.6° to 30.2° as the normal stress increased from 19.5 to 76.7 kPa. Negligible shearing rate effects were observed for the internal shear strength of this material. The TDA-concrete interface failure envelope is linear with a friction angle of 22.6°. The dilation angle decreased with increasing normal stress for the TDA internal shear tests, whereas only contraction was observed for the TDA-concrete interface shear tests. Displacements at failure for the TDA internal shear tests ranged from 333 to 439 mm, and were two to three times larger than those for the TDA-concrete interface shear tests.

Although tire-derived aggregate (TDA) has been used widely as lightweight fill in civil engineering applications, the properties governing its response under cyclic loading are not well understood. Reliable data on the evolution of shear modulus and damping ratio with cyclic shear strain amplitude are needed for the prediction of the seismic response of TDA fills, especially those with larger particle sizes up to 300 mm (Type B TDA). This study presents the results of cyclic simple shear tests performed on Type B TDA using a new large-scale testing device for vertical stresses ranging from 19.3 to 76.6 kPa and shear strain amplitudes ranging from 0.1 to 10%. The shear modulus of Type B TDA has a maximum value of 2,386 kPa and decreases with increasing shear strain amplitude, which is smaller in magnitude and similar in trend to natural granular soils in this vertical stress range. Continuous volumetric contraction was observed during cyclic loading for all stress levels. The damping ratio for Type B TDA showed a different behavior from granular soils, with a relatively high magnitude of 2–26.8% at the lowest shear strain amplitude (0.1%), followed by a decreasing–increasing trend with increasing amplitude. The shear modulus was found to follow a power-law relationship with vertical stress, similar to granular soils, and the damping ratio was not sensitive to vertical stress level.

The evolution of cyclic resistance with multiple loading events is evaluated for nonplastic and low-plasticity silts with plasticity indices of 0 and 6, respectively. A series of direct simple shear tests with multiple cyclic loading and reconsolidation stages is performed on young slurry-sedimented specimens. Evolution of cyclic strength with a series of multiple loading events is examined with respect to densification from postcyclic reconsolidation, shear strain–induced fabric, and initial consolidation history. Initially normally consolidated specimens of both silts develop progressive increases in cyclic strength with prior strain history, with cyclic resistance ratios ultimately exceeding 0.6. Specimens consolidated with an initial overconsolidation ratio of 2 experience an 18–32% loss of cyclic strength following the first stage of cyclic shearing and reconsolidation. The two silts develop similar magnitudes of reconsolidation strain, but the low-plasticity silt requires more volumetric strain (and therefore more loading events) to develop large cyclic strengths. Implications for future advances in liquefaction triggering correlations and engineering practice are discussed.

Particulate media filters and infiltrometers play a critical role in separation of solids and liquids in infrastructure applications that require drainage and transmission of water or other liquids. Although the primary mechanisms of particle removal within a filter are well developed (straining, attachment, and detachment), the impact of geochemistry on fine-grained particle removal is not well developed, especially in terms of the impact of clustering of clay particles as a function of ionic strength. This experimental investigation quantifies the impact of ionic strength, median filter-grain diameter, shape of filter grain, and inflow concentration of clay suspensions on the reduction of saturated hydraulic conductivity ($K_s$) in a sand column. Additionally, three first-order reaction coefficients for each condition are optimized by nonlinear least-squares fit to the data to account for straining, attachment, and detachment mechanisms for the clay deposition. Georgia kaolinite was tested as a representative clay material, and the experimental retention profiles were determined as a function of depth. Results demonstrate that the amount of strained kaolinite decreases as the ionic concentration of the clay suspension increases, which results from the formation of small clay clusters at relatively high ionic strength. Furthermore, the observed reduction in $K_s$ of the sand-filter medium is attributable primarily to straining in the first 7 cm of the depth of the filter. Decreasing the ionic strength of the clay suspension results in increased size of the clay clusters, which leads to flocculation of the clay and a relatively large size ratio between the kaolinite clusters and sand particles ($d_c/d_s$). This produces a significant reduction of $K_s$ as a result of straining at the surface of the filter. The work demonstrates that ionic strength, clay particle-cluster size, and filter sand-grain size must be considered as major factors in particle retention and reduction of hydraulic conductivity within a filter or infiltration trench.

The installation of stone columns is widely adopted to prevent liquefaction. The main improvement mechanisms of stone columns as a liquefaction countermeasure are drainage, stiffening, and densification. This paper investigates the effectiveness of stone columns as a liquefaction remediation. Twenty-four case studies, in which SPT and CPT tests are performed before and after stone column reinforcement, are used as a basis of the research. Densification and stiffening mechanisms are investigated and their individual and combined effects are analyzed. Comparative results indicate that considering the densification and stiffening effects considerably improve the assessment of liquefaction potential of reinforced soil by stone columns.